Inkscore watches how writers write — not just what they write. Behavioral analysis scores every article for authentic human authorship.
Inkscore Report
Binary AI/human labels with high false-positive rates. Your writers get flagged unfairly. Your clients lose trust in the process.
Running every article through a third-party detector after it's written means you're auditing output, not verifying process.
One AI-generated article that slips through damages client relationships. Content agencies need a system, not a spot-check.
Process-level verification in three steps.
Writers compose in Inkscore's distraction-free editor. Behavioral telemetry silently captures keystroke patterns, editing flow, and composition rhythm.
On submission, weighted behavioral signals produce a 100-point Inkscore. Pure behavioral math — zero AI involved in the analysis.
Editors see the full score breakdown, paste logs, and session details. Approve, flag, or request revision — all in one streamlined workflow.
From the writing canvas to the review dashboard — every step is designed for clarity.
Deliver verifiable proof reports with every article. Differentiate your agency with a process guarantee that builds client confidence.
Review submissions with full scoring context — paste logs, session metadata, and behavioral signals. Approve, flag, or request revisions with inline comments.
Write in a beautiful, distraction-free editor with sprint tracking and writing themes. Your Inkscore builds over time as the system learns your natural patterns.
Inkscore captures behavioral patterns, not content. Here's how we protect your team's data.
AI detectors analyze the output text and guess whether it was AI-generated. Inkscore is fundamentally different — it watches the writing process in real time, analyzing behavioral signals across four scoring categories: Composition, Rhythm, Consistency, and Identity. We verify the human, not the text.
Inkscore captures timing data — keystroke intervals, pause patterns, deletion rates, cursor movement, and paste events. It does not record the actual words typed, take screenshots, or perform any screen recording. Think of it as a fitness tracker for writing behavior.
Inkscore's behavioral signals specifically detect this. The system tracks whether content was built incrementally through natural editing or appeared in suspiciously large chunks. Combined with behavioral baseline comparisons, surface-level workarounds are reliably caught.
Legitimate pastes — quotes, statistics, URLs — are expected in professional writing. Inkscore's scoring algorithm considers paste size, frequency, clustering, and crucially whether pasted content was subsequently edited. A few small, edited pastes barely affect the score.
Every article starts at 100. Behavioral signals across four categories — Composition, Rhythm, Consistency, and Identity — can deduct points based on suspicious patterns. The final Inkscore reflects overall writing authenticity. Scores above 80 are green, 60-80 amber, below 60 red.
No. The scoring algorithm is entirely deterministic — pure math based on behavioral signals. There is no machine learning model, no LLM, and no AI involved in computing an Inkscore. This makes the results fully reproducible and auditable.
Writers are informed about telemetry collection and must acknowledge a data notice before their first writing session. All behavioral data is encrypted, isolated per organization, and never shared with third parties. Writers can view their own scores and trends at any time.
Join the content agencies that trust behavioral verification over AI detection.